{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adversary-detection-in-neural-networks-via","title":"Adversary Detection in Neural Networks via Persistent Homology","arxiv_id":"1711.10056","date":"2017-11-28","proceeding":null,"authors":["Thomas Gebhart","Paul Schrater"],"abstract":"We outline a detection method for adversarial inputs to deep neural networks.\nBy viewing neural network computations as graphs upon which information flows\nfrom input space to out- put distribution, we compare the differences in graphs\ninduced by different inputs. Specifically, by applying persistent homology to\nthese induced graphs, we observe that the structure of the most persistent\nsubgraphs which generate the first homology group differ between adversarial\nand unperturbed inputs. Based on this observation, we build a detection\nalgorithm that depends only on the topological information extracted during\ntraining. We test our algorithm on MNIST and achieve 98% detection adversary\naccuracy with F1-score 0.98.","url_abs":"http://arxiv.org/abs/1711.10056v1","url_pdf":"http://arxiv.org/pdf/1711.10056v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adversary-detection-in-neural-networks-via","repo_url":"https://github.com/tgebhart/tf_activation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1711.10056","atlas_url":"https://app.syntology.ai/?focus=1711.10056","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}